National greenhouse gas emissions baseline scenarios learning from experiences in developing countries

National Greenhouse Gas
Emissions Baseline Scenarios
Learning from Experiences in
Developing Countries

A report by the Danish Energy Agency, the Organisation for
Economic Co-operation and Development and the UNEP Risø
Centre, based on contributions from experts in Brazil, China,
Ethiopia, India, Indonesia, Kenya, Mexico, South Africa, Thailand
and Vietnam


2

Foreword
Successful policy-making hinges on robust analysis
of expected future developments. Planning for climate
change policy is no exception: understanding likely future
trends in greenhouse-gas emissions is important not only
for domestic policy-making but also for informing countries’ positions in international negotiations on climate
change. To this end, many countries have developed
scenarios describing plausible future trends in emissions.
Generally, the most important among these scenarios is
the baseline or business-as-usual scenario, which aims to
characterise future emissions on the assumption that no
new climate change policies will be adopted.
Greenhouse gases are emitted as a result of many
different types of economic activity. As a result, preparing emissions scenarios involves making decisions and
assumptions concerning many different underlying drivers
of emissions, ranging from political factors to the type of
modelling tools used. Such decisions are often governed

by constraints on resources, including skills, information
and funding. Naturally, these constraints, and how they
affect climate change policy-making, vary from country to
country.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

It is not surprising, therefore, that existing approaches
to developing national baseline scenarios are highly
disparate. Yet this diversity is increasingly at odds with
developments in the international negotiations under
the United Nations Framework Convention on Climate
Change. Since 2011, emissions reduction pledges put
forward by Parties are formally recognised under the
Convention. Some Parties have pledged quantified emissions reductions and actions for 2020 relative to their
baseline scenario. This means that the expected magnitude of the overall global mitigation effort and, hence,
the likelihood of achieving the agreed goal of limiting
global warming to 2°C, depends in part on the way those
baseline scenarios are calculated. Consequently, improving international understanding of those scenarios and
achieving a minimum level of comparability is important.

Kristian Møller
Deputy Director General,
Danish Energy Agency

3

While perhaps desirable from the point of view of the
international climate change regime, the establishment of

universally-applicable guidelines for developing baseline
scenarios is likely to be technically difficult and politically
challenging. Given these constraints, this report aims
rather to contribute to a better understanding of the
issues and challenges involved in drawing up baseline
scenarios, by documenting and drawing lessons from the
breadth of existing practices in a range of countries. This
existing diversity is both a key asset for gradually increasing the robustness of baseline scenarios, but also the
reason for a lack of comparability. We hope that this work
shows the value of improving transparency in baseline
scenarios and we invite governments and other stakeholders to continue to share experiences in this area.

Simon Upton
Director, OECD
Environment Directorate

John Christensen
Head, UNEP Risø Centre


4

Acknowledgements
This publication has been made possible thanks to significant in-kind contributions from experts in ten developing
countries – Brazil, China, Ethiopia, India, Indonesia,
Kenya, Mexico, South Africa, Thailand and Vietnam –
who were willing to share their experiences in establishing
national baseline emissions scenarios at seminars and
workshops and by writing up the reports included in Part
2 of this publication.

Sincere thanks go to the authors of the country
contributions:
• Brazil: Emilio Lèbre La Rovere (Professor, Energy and
Environmental Planning, at COPPE/UFRJ - Institute
of Graduate Studies and Research in Engineering,
Federal University of Rio de Janeiro).
• China: Liu Qiang and Jiang Kejun (Energy Research
Institute, ERI, National Development and Reform
Commission).
• Ethiopia: Wondwossen Sintayehu
Wondemagegnehu (Environmental Protection
Agency).
• India: Atul Kumar and Ritu Mathur (The Energy and
Resources Institute, TERI).
• Indonesia: Syamsidar Thamrin (National Planning
Agency, Bappenas).

• Kenya: Fatuma M. Hussein (Ministry of Environment
and Mineral Resources).
• Mexico: Lucía Cortina Correa and Iliana Cárdenes
(Ministry of Environment and Natural Resources,
Semarnat).
• South Africa: Thapelo Letete, Harald Winkler, Bruno
Merven, Alison Hughes and Andrew Marquard
(Energy Research Centre, ERC, University of Cape
Town).
• Thailand: Chaiwat Muncharoen (Thailand Greenhouse
Gas Management Organisation).
• Vietnam: Tran Thuc, Huynh Thi Lan Huong and Dao
Minh Trang (Institute of Meteorology, Hydrology and

Environment in Vietnam).


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

We would also like to thank Liz Stanton (formerly
Stockholm Environment Institute, now Synapse Energy)
who contributed at different stages of the publication
process. Further, we are very grateful for the valuable
comments received from the following reviewers: Alexa
Kleystueber (Chile Ministry of Environment); Marta
Torres Gunfaus (ERC, University of Cape Town and
Mitigation Action Plans and Scenarios [MAPS] project);
Kiyoto Tanabe (Institute for Global Environmental
Strategies, Japan); Jane Ellis (the Organisation for
Economic Co-operation and Development, OECD); Katia
Simeonova, Sylvie Marchand and Babara Muik (United
Nations Framework Convention on Climate Change,
UNFCCC); Charlie Heaps (Stockholm Environment
Institute); Christa Clapp (Thomson Reuters Point
Carbon); Todd Ngara and Jørgen Fenhann (UNEP
Risø Centre); and Kelly Levin, David Rich and Jared
Finnegan (World Resources Institute). Reviewers commented on the draft report in their respective personal capacities. Trevor Morgan (Menecon Consulting) reviewed
and edited the final draft of Part 1 of the report. Language
revisions in Part 2 of the report were made by Josephine
Baschiribod.
Jacob Krog Søbygaard, Peter Larsen, Sixten Rygner
Holm and Ulla Blatt Bendtsen (all Danish Energy
Agency), Andrew Prag (OECD) and Daniel Puig (UNEP
Risø Centre) wrote Part 1 of this report. The Danish

Energy Agency, the OECD and the UNEP Risø Centre
provided financial and in-kind contributions for this work.

Copyright © 2013: The Danish Energy Agency (DEA),
the Organisation for Economic Co-operation and
Development (OECD) and the UNEP Risø Centre (URC).
This publication may be reproduced in whole or in part
and in any form for educational or non-profit purposes
without special permission from the copyright holder,
provided acknowledgement of the source is made. DEA,
OECD and URC would appreciate receiving a copy of any
publication that uses this publication as a source. No use
of this publication may be made for sale or for any other
commercial purpose whatsoever without prior permission
in writing from DEA, OECD and URC.
Disclaimer
The designations employed and the presentation of the
material in this publication do not imply the expression of
any opinion whatsoever on the part of DEA, URC, OECD
or OECD member countries concerning the legal status
of any country, territory, city or area or of its authorities,
or concerning delimitation of its frontiers or boundaries. Moreover, the views expressed do not necessarily
represent the decision or the stated policy of DEA, URC,
OECD or any OECD member country, nor does citing
of trade names or commercial processes constitute
endorsement.
ISBN (printed version): 978-87-7844-989-4
ISBN (online version): www 978-87-7844-987-0
April 2013

Contact e-mail address:

5


6

Table of Contents

Foreword .......................................................................2
Acknowledgements .......................................................4
Key terminology .............................................................8
Acronyms ......................................................................9
Main findings ...............................................................10

Part 1: Synthesis report
Chapter 1: Introduction .............................................14
Role of baseline scenarios ...........................................16
Relevant existing literature ...........................................17
Related initiatives .........................................................17
Structure of the report .................................................17

Chapter 4: Data management ...................................38
Emissions inventories ..................................................38
Socio-economic data and emissions factors................40
Institutional arrangements
and capacity constraints ..............................................42

Chapter 2: Model choice and use.............................18
Types of models .........................................................18

Existing versus purpose-made models ........................22
Land-use sector emissions modelling ..........................23
Institutional arrangements and capacity constraints .....24

Chapter 5: Transparency and inclusiveness
in developing baseline scenarios .........................44
Stakeholder involvement ..............................................45
Peer review..................................................................46
Comparing in-country and
supra-national model projections .............................47

Chapter 3: Assumptions and sensitivity analyses ...26
Definition and purpose .................................................26
Existing versus additional policies ................................28
Exclusion criteria..........................................................29
Base year ....................................................................30
Revisions .....................................................................31
Key drivers ..................................................................32
Technology development and learning .........................34
Sensitivity analyses ......................................................35
Comparing baselines ...................................................36

Chapter 6: Reflections on key aspects
of developing a baseline scenario ............................48
Transparency in baseline setting ..................................48
Key defining factors in baseline scenarios ....................49
Uncertainty in baseline scenarios .................................49
Towards elements of ‘good practice’ ...........................49

National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

Part 2: Country Contributions
Brazil (UFRJ) ..............................................................54
China (ERI) .................................................................66
Ethiopia ......................................................................74
India (TERI) ................................................................80
Indonesia ...................................................................92
Kenya .......................................................................108
Mexico .....................................................................114
South Africa (ERC) ...................................................124
Thailand ...................................................................138
Vietnam ....................................................................144

Appendix: Background information
About us....................................................................154
The Baseline Work Stream ........................................154

7


8

Key terminology
Base year: An historical year which marks the transition
from emissions estimates based on an inventory to modelling-based estimates of emissions volumes. In many
countries the base year coincides with the latest year for
which emissions inventory data are available. In other
instances, there may be a gap of a few years between
the latest year for which inventory data are available and

the initial year for which projections are made.
Exclusion criteria: A sub-set of assumptions concerning
policies or technologies which, while feasible in principle,
are ruled out on ideological or economic grounds.
Existing policies: Existing policies are those that have
been legally adopted by a certain cut-off date. Some policies that have been implemented before the cut-off date
may have had an impact on emissions before that date,
while others may only have an impact later on.
Forecast: A projection to which a high likelihood is
attached.
Model: A schematic (mathematical, computer-based)
description of a system that accounts for its known or
inferred properties. The terms ‘model’ and ‘modelling
tool’ are used interchangeably in this publication.

Projection: Estimates of future values for individual parameters, notably those that are key drivers of emissions
in a scenario.
Reference year: Year against which emissions reduction pledges are measured. This could be a past year
(for example, 1990 in the case of the European Union’s
commitment under the Kyoto Protocol) or a future year
(as is the case for those non-Annex I countries that have
defined their pledge relative to a baseline scenario).
Scenario: A coherent, internally consistent and plausible
description of a possible future state of the world given
a pre-established set of assumptions. Several scenarios
can be adopted to reflect, as well as possible, the range
of uncertainty in those assumptions.
• Baseline scenario: A scenario that describes future
greenhouse-gas emissions levels in the absence of
future, additional mitigation efforts and policies. The

term is often used interchangeably with business-asusual scenario and reference scenario.
• Mitigation scenario: A scenario that describes future
emissions levels taking account of a specified set of
future, additional mitigation efforts and policies.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

9

Acronyms
BaU: Business-as-Usual
CCXG: Climate Change Expert Group (a group of
government delegates and experts from OECD and other
industrialised countries)
CETA: Carbon Emissions Trajectory Assessment (a
model)
CGE: Computable General Equilibrium (a type of model)
CO2e: Carbon dioxide equivalent (a unit of measurement)

MAPS: Mitigation Action Plans and Scenarios (a multicountry programme)
MARKAL/TIMES: MARKet ALlocation / The Integrated
Markal/Efom System (a model in its first – MARKAL – and
second – TIMES – generation versions)
MEDEE: Long-term Demand Prospective Model
MESSAGE: Model for Energy Supply Strategy
Alternatives and their General Environmental impact

COMAP: Comprehensive Mitigation Assessment Process
(a model)

MW: Megawatt

COP: Conference of the Parties to the United Nations
Framework Convention on Climate Change

NEMS: National Energy Modelling System (an economic
and energy model)

DEA: Danish Energy Agency

NGO: Non-Governmental Organisation

EFOM: Energy Flow Optimisation Model

OECD: Organisation for Economic Co-operation and
Development

ERC: Energy Research Centre (University of Cape Town,
South Africa)

NAMAs: Nationally Appropriate Mitigation Actions

ERI: Energy Research Institute (China)

POLES: Prospective Outlook on Long-term Energy
Systems (a model)

GDP: Gross Domestic Product

PPP: Purchaising Power Parities

GHG: Greenhouse Gas

REDD: Reduced Emissions from Deforestation and forest
Degradation

Gt: Gigatonne
GW: Gigawatt

RESGEN: Regional Energy Scenario Generator Module
(a model)

IEA: International Energy Agency

SGM: Second Generation Model

IPAC: Integrated Policy Model for China

TERI: The Energy and Resources Institute (India)

IPCC: Intergovernmental Panel on Climate Change

UFRJ: Federal University of Rio de Janeiro

LEAP: Long-range Energy Alternative Planning System (a
modelling framework)

UN: United Nations

LULUCF: Land Use, Land Use Change and Forestry
LUWES: Land Use Planning for loW Emissions development Strategy (a decision support tool)
MAC: Marginal Abatement Cost
MAED: Model for Analysis of Energy Demand

UNEP: United Nations Environment Programme
UNFCCC: United Nations Framework Convention on
Climate Change
URC: UNEP Risø Centre
WEM: World Energy Model


10

Main findings
The following summary highlights the key findings of the
main content of Part 1, Chapters 1-5. The authors’ reflections on good practice for baseline setting can be found
in Chapter 6 and are not summarised here. Throughout
the document, mention of national experiences refers
only to the ten countries contributing to this publication.

Chapter 1: Introduction
• A national emissions baseline scenario aims to inform
decision makers about how greenhouse-gas (GHG)
emissions are likely to develop over time under certain given conditions. Even if developed primarily for
national policy-planning purposes, baselines can also
be important in an international context.
• Within the context of the international climate change
negotiations, some developing countries have defined
their mitigation actions on the basis of deviations from

their baseline scenarios. Five of the ten participating
countries – Brazil, Indonesia, Mexico, South Africa and
Vietnam – fall into this category. In these countries, the
model and assumptions behind the baseline affect the
resulting targeted emissions reduction levels, making these baselines particularly important for climate
change negotiations.

• For all developed and developing countries (irrespective of the type of pledge), baseline scenarios are
valuable for planning purposes, including to support
the design of energy and climate change policy and
investment decisions.
• There is currently no international guidance on how to
develop baseline emissions scenarios and there is no
explicit requirement for developing countries to report
on emissions baselines.
• The ten countries differ widely in their sources of GHG
emissions. For some countries, the energy sector is
the most important emissions sector, while for others the land-use sector and/or the agricultural sector
dominates the emissions picture.

Chapter 2: Model choice and use
• The choice of modelling tool used to prepare baseline
scenarios tends to be driven by a trade-off between
performance (in the form of sophistication and anticipated accuracy) and resources available (including
human capacities and data availability). Familiarity
with the tool, ease-of-use and financial and technical
assistance from other, more experienced countries,


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

all contribute to shaping decisions on model choice.
In general, resource constraints often play a dominant
role in model selection in the participating countries.
• To model energy sector emissions, most participating
countries rely on bottom-up models, which provide
a fairly detailed representation of the energy system,
albeit at the expense of a more complete representation of macroeconomic trends and feedbacks. Few
countries use simple extrapolation top-down models.
Hybrid models can combine elements of top-down
and bottom-up models to overcome the limitations
of both types, but are often complex to build. The
onerous requirements of hybrid models, in terms of
both data and expertise, seem to make them difficult
to apply in most countries; at the moment, only China,
India and South Africa, among the ten participating
countries, use them.
• In general, most countries use existing models to
develop their baseline scenarios. One reason for this
is that developing a model from scratch is demanding
and resource-intensive, and there is no guarantee that
the model will be better than an existing alternative.
Some countries tailor existing tools to satisfy their
specific needs. Mexico previously used a fully purposemade model.
• One might expect that countries whose land-use sector emissions account for a large proportion of national
emissions would have a stronger interest in investing
in building modelling capacity in this area. However,
experience suggests that availability of existing tools
and processes, as well as resource constraints, are the
main determinants of the sophistication of the modelling approach used. One reason for this may be the

inherent uncertainty that charcaterises the modelling of
emissons from the land-use sector: beyond a certain
level of complexity, the incremental effort needed to
enhance the output appears to be significant.
• Baseline scenarios are not an end in themselves:
they support broader national and often international
processes. As a result, the process of setting baseline
scenarios is inevitably governed by the institutional
arrangements put in place to implement those broader
processes. These arrangements may have been
designed with other purposes in mind and so may not

11

be best adapted to the task of preparing a baseline
scenario. Increased awareness about the importance
of baselines, coupled with stronger political mandates,
and increased experience and resources, could help
improve governance arrangements and enhance interagency cooperation.

Chapter 3: Assumptions and sensitivity analyses
• There is no commonly-agreed definition of baseline
scenario. It is defined in this report as “a scenario that
describes future greenhouse-gas emissions levels in
the absence of future, additional mitigation efforts and
policies”. In principle this could include either scenarios
that eliminate effects of all climate policies or scenarios
that model effects of existing climate policies (but in
both cases excluding possible future policies). Which
policies are considered ‘existing’ can have a great

impact on the resulting emissions baseline scenario.
• Most countries include the estimated effects of some
existing policies in their baselines. The selection of
which policies to include is not necessarily restricted to
climate change policies, because policies implemented
on grounds other than climate change mitigation can
have an impact on emissions levels. Worth noting is
South Africa’s choice to develop two baseline scenarios – one with existing policies and a second, no-policy
scenario. The government of South Africa adopted the
latter as its official baseline (using a range, rather than
a single point estimate for each year).
• How to select ‘existing policies’ and how to model the
impacts of any one approach (‘no policies’ or ‘only
existing policies’) are key questions, in that the choices
made greatly influence the results of the analysis.
Given the wide range of possible answers to these
questions, combined with the lack of commonlyagreed approaches in this area, clarity on the steps
taken in the analysis will be crucial to understand the
meaning of baseline scenarios.
• Exclusion criteria are a sub-set of assumptions concerning policies or technologies which, while in principle feasible, are ruled out on ideological or economic
grounds. Implicitly or explicitly, all countries introduce
exclusion criteria in their baselines. For example, cost


12

minimisation is central to the modelling approach
used in India and South Africa. Baseline scenarios
seldom depart from established technologies and often
introduce cost constraints, which are in themselves

exclusion criteria.

sector emissions estimates can be high, sensitivity
analyses have not been used to estimate the resulting
potential impacts on baseline scenarios.

Chapter 4: Data management
• The choice of base year (or start year) for the baseline
scenario depends on both technical and political considerations. Agreement on which criteria are to guide
the choice of base year could be helpful, recognising
that there can be valid reasons for choosing different
base years in different countries. Choosing a year in
which emissions in the country departed from the
trend in previous years can mask the likely evolution of
emissions in the future.
• Only one participating country (Mexico) has made legal
provisions for regularly revising the baseline scenarios
as well as mitigation trajectories. Those provisions
specify a time period for revision and update and
define circumstances that may trigger a more frequent
review.
• Key modelling assumptions regarding socio-economic
and other factors driving projections may be politicallydetermined. Among the most critical assumptions are
estimated changes in gross domestic product (GDP),
population, energy prices and the sectoral composition
of national income. For some countries, these assumptions are based on government targets, notably GDP
targets. However, these assumptions may not always
correspond to ’the most likely’ outcome.
• Most countries use national data sources for key
drivers such as GDP, population and energy prices,

rather than datasets available internationally (from, for
example, the United Nations Population Division, the
World Bank, the OECD or the IEA).
• Sensitivity analyses assess the uncertainty of the output of a model with respect to its inputs, thus providing an indication of the robustness of model outputs.
Generally, the extent of sensitivity analyses carried out
to date has been limited, though baseline developers
do recognise the importance of sensitivity analysis.
Sensitivity analysis for GDP growth assumptions is
critical (especially for some sectors) and deserves
special scrutiny. Further, while uncertainty of land-use

• Data management issues are important for many
aspects of baseline-scenario development, as is the
completeness of the national emissions inventory. In
addition to problems with basic data availability, a key
challenge is to reconcile existing data collection frameworks with the IPCC source categories. If data are
unavailable, scenarios must rely on assumed growth
trends.
• The accuracy of emissions factors used in baseline
calculations differs greatly among countries. Given
the difficulty of calculating country-specific emissions
factors for all sectors, many countries use default IPCC
emissions factors. In countries such as Brazil, with long
experience of emissions modelling, country-specific
emissions factors are used. In other countries, countryspecific emissions factors are often developed only
for certain high-emissions sectors (as is the case in
Vietnam and Thailand, for example). Preparing countryspecific emissions factors is a resource-intensive task.
• The inventory included in a country’s most recent
national communication to the UNFCCC may not
contain the latest data available (as countries may

update their inventory more regularly than they report
to the UNFCCC). In some baseline scenarios, the base
year coincides with the latest year for which emissions
inventory data are available; in other cases, the base
year itself is modelled. In the latter case, countries are
in effect estimating emissions levels for that base year.
How well this can be done depends on the quality
of historical emissions data. Clarity on the approach
taken is crucial for understanding the baseline scenario
• Several of the participating countries have established
a coordinating committee or working group to organise
and allocate the inter-agency work related to national
climate change mitigation policies. Besides fulfilling
an administrative role, such a framework can help to
ensure political support in the different governmental agencies. Without this, the lack of international


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

guidance on baseline-setting means that it is left to
resource-constrained government agencies to decide
on the myriad options involved in baseline development, often in the absence of a coherent overview.
• Data management presents a challenge for most participating countries. Chief amongst those challenges is
lack of high quality data. Improving data accuracy represents an ongoing concern for most countries; some
countries rely on international assistance to improve
practices and standards.

Chapter 5: Transparency and inclusiveness in
baseline setting
• Although not all countries state transparency and

international credibility as specific objectives when
setting a baseline, there is broad acknowledgement
among the participating countries that these are key
concerns. Accordingly, in the process of developing
their baseline, countries have made available varying
levels of information regarding the assumptions chosen
for the preparation of the baseline.
• Countries have had varying experiences with stakeholder consultation in the baseline development
process, including the extent to which stakeholders
(notably in industry, civil society, labour and government) are consulted and at which stage in the process.
The stakeholder-consultation process conducted in
South Africa during the preparation of its Long Term

13

Mitigation Scenarios was particularly comprehensive. Mexico is planning an extensive stakeholder
consultation.
• International review of national baselines can be a
politically sensitive matter. Informal peer reviews can be
one way around this difficulty. By increasing transparency, peer review can add to both the robustness and
credibility of the baseline. South Africa is the first of the
participating countries to have conducted this type of
peer review.
• Some participating countries note that there are benefits from comparing and understanding differences
across various studies on baselines for the same country, whether they are domestic or international studies.
For example, the government of India commissioned
five different baseline studies, to benefit from the different approaches each study followed.
• International peer review can be particularly beneficial
when it is conducted in an open manner, with participating parties having access to each other’s data and
models. Besides, analysing a national baseline against

an international background can shed new light on
key international developments of relevance to that
national baseline (for example, it can help understand
the sensitivity in demand for fossil fuels due to changes
in GDP in different regions).


14

Chapter 1: Introduction
This report reviews national approaches to preparing
baseline scenarios of greenhouse-gas (GHG) emissions.
It does so by describing and comparing in non-technical
language existing practices and choices made by ten
developing countries – Brazil, China, Ethiopia, India,
Indonesia, Kenya, Mexico, South Africa, Thailand and
Vietnam. The review focuses on a number of key elements, including model choices, transparency considerations, choices about underlying assumptions and challenges associated with data management. The aim is to
improve overall understanding of baseline scenarios and
facilitate their use for policy-making in developing countries more broadly.1

The findings are based on the results of a collaborative
project involving a number of activities undertaken by the
Danish Energy Agency, the Organisation for Economic
Co-operation and Development (OECD) and the UNEP
Risø Centre (URC), including a series of workshops on
the subject (Box 1). The ten contributing countries account for approximately 40% of current global GHG emissions2 – a share that is expected to increase in the future.
The breakdown of emissions by sector varies widely
among these countries (Figure 1). In some countries,
the energy sector is the leading source of emissions; for
others, the land-use sector and/or agricultural sector

dominate emissions.
The report underscores some common technical and
financial capacity gaps faced by developing countries
when preparing baseline scenarios. It does not endeavour to propose guidelines for preparing baseline scenarios. Rather, it is hoped that the report will inform any
future attempts at preparing such kind of guidelines.

1. This report does not cover project or sector-level baselines (for example, for a project to recover methane from landfills, or to increase the use of
renewable energy for electricity generation), which are common to offset-based carbon markets.
2. Based on total GHG emissions in 2010 as estimated in the IEA’s World Energy Outlook 2012.


-20

National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

15

-40

-130
-60

Brazil

(2005)

China
(2005)

Ethiopia

(1995)

India

(2000)

Industrial Processes

Energy

Endonesia
(2000)

Kenya
(1994)

Agriculture

Mexico
(2006)

LULUCF

South
Africa
(2000)

Thailand
(2000)

Vietnam
(2000)

Waste

Figure 1: Emissions and sinks in participating countries
100
80
60
40
20
0
-20
-40

-130
-60

Brazil

(2005)

China
(2005)

Energy

Ethiopia
(1995)

India

(2000)

Industrial Processes

Indonesia
(2000)

Kenya
(1994)

Agriculture

Mexico
(2006)

Land-use

South
Africa
(2000)

Thailand
(2000)

Vietnam
(2000)

Waste

Note: This figure is indexed to highlight the different emissions compositions in the ten countries. The indexation is
done by setting the sum of emissions (excluding sinks) to 100. The differences in absolute size in emissions across the
countries are not visible here.
Source: National Communications to the UNFCCC.

Box 1
Origins of this report
In 2011, the DEA invited five developing countries –
Ethiopia, Kenya, Mexico, South Africa and Vietnam – to
share information on how they had prepared their national
GHG emissions baseline scenarios. At the same time,
the OECD was working on the development of baseline
scenarios under the aegis of the Climate Change Expert
Group (CCXG).
It was decided to bring these two activities together by
organising a series of workshops in 2011 and 2012. The

UNEP Risø Centre joined the collaborative project at this
point, to provide additional technical expertise. As the
workshops progressed, experts from five other countries
– Brazil, China, India, Indonesia and Vietnam – joined
the project, bringing the final list of participating countries to ten. The countries shared existing practices and
challenges they have faced in establishing their baseline
scenarios. More background information about the collaboration can be found in the appendix.


16

The report does not address practices in developed

countries. However, some of the participating countries
suggested that future work on best practices in preparing national baseline scenarios should take into account
experience in developed countries as well.

Change (UNFCCC), as some developing countries have
defined their mitigation pledges in terms of reductions
from their respective baselines. As a result, the strength
of overall efforts to reach the internationally-agreed mitigation target of limiting global warming to 2°C is indirectly
linked to the reliability of national baseline scenarios.4

Role of baseline scenarios

Against this background, there is growing interest in
both understanding and improving approaches to
calculating baseline scenarios. There is little guidance
available to aid this process, particularly for developing
countries. Guidelines exist for the preparation of National
Communications by parties to the UNFCCC, as well as
for compiling the forthcoming biennial update reports
(Box 2). However, no specific guidelines or protocols are
available to assist countries in preparing their national
baseline scenarios.

We define baseline scenario as a scenario that describes
future GHG emissions levels in the absence of future, additional mitigation efforts and policies.3 Baseline scenarios
are used routinely to support domestic policy planning
as well as to inform national positions in international
climate-change negotiations. In recent years national
baselines have grown in importance in the context of
the United Nations Framework Convention on Climate

Box 2
UNFCCC guidelines relevant for
reporting by non-Annex I parties
Guidelines for national communications
(Decision 17/CP.8)
• Protocols for the compilation of national GHG inventories, including inventory year, tier methods, default
emissions factors, activity data, key category analysis
and sectoral approaches, gases and global warming
potentials.
• Protocols for describing programmes containing measures to mitigate climate change.
Guidelines for biennial update reports
(Decision 2/CP.17)
• Protocols for the compilation of the national GHG
inventory report.

• Protocols for describing mitigation actions, including
quantitative goals; methodologies and assumptions;
objectives of the actions; progress of implementation; information on international market mechanisms;
monitoring, reporting and verification arrangements;
financial, technology and capacity-building needs; and
support received.
In addition, the sixth compilation and synthesis of national communications from non-Annex I parties to the
UNFCCC (FCCC/SBI/2005/18/Add.3) includes information about expected GHG abatement, mitigation opportunities, examples of measures implemented or planned
by developing countries and indications of the financial
resources required to implement identified measures or
projects.

Source: presentation by Dominique Revet (UNFCCC Secretariat) at a side event held in Bonn on 15th May 2012.

3. See the Key Terminology section at the front of this report for more detail on this and related terms.
4. A similar case could be made for so-called nationally appropriate mitigation actions (NAMAs). This is because NAMAs are often prioritised by
means of the same tools used for preparing baseline and mitigation scenarios. Given that, in some instances, bilateral or multi-lateral funding
sources are sought to finance NAMAs, clarity on approaches to scenario development could facilitate funding agreements.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

Relevant existing literature
Preparing baseline emissions scenarios invariably involves
the use of energy and emissions modelling techniques.
For many years, researchers, governments and international organisations have been working to develop and
improve these techniques. This report does not aim to
provide a comprehensive overview of the subject, so a
full academic literature review is not included. Few reports
have focused specifically on national baseline scenario
development. Some relevant works include:
• In-depth reviews on national communications, by the
UNFCCC secretariat.5
• Greenhouse gas emission projections and estimates of
the effects of measures: moving towards good practice. A 1998 OECD information paper aimed to identify
good practices in the preparation of greenhouse-gas
emissions projections in Annex I countries.6
• Projecting Emissions Baselines for National Climate
Policy: Options for Guidance to Improve Transparency,
by C. Clapp and A. Prag. A 2012 OECD/IEA information paper providing options and elements for guidance and potential future guidelines in baseline development (published under the CCXG).7
• Developing Baselines for Climate Policy Analysis, by E.
A. Stanton and F. Ackerman. A 2011 UNEP document
prepared as a part of an initiative aimed to support
long-term planning for climate change, which included

guidance on baseline scenario development.8

Related initiatives
Complementing the work leading to this report, two other
international initiatives may be of interest to countries
seeking to improve how they go about preparing their
baseline scenario:
• The Mitigation Action Plans and Scenarios (MAPS)
programme. This programme aims to share best

17

practices on low-carbon transition planning and scenario development, including preparing baseline scenarios. It is a collaborative effort involving developing
countries, led by the University of Cape Town’s Energy
Research Centre in partnership with SouthSouthNorth,
a network organisation. The programme is active in
five Latin American countries: Argentina, Brazil, Chile,
Colombia and Peru.9
• The Mitigation Accounting Initiative. Launched by
the World Resources Institute in 2012, this multi-stakeholder initiative seeks to develop voluntary guidelines
to increase the consistency and transparency with
which a wide array of stakeholders, including governments, account for GHG reductions arising from
specific mitigation actions and goals. These guidelines
include recommendations for developing baseline
scenarios.10
While both initiatives are dealing with baseline scenarios,
it is not their exclusive focus. Furthermore, a number
of other initiatives are also relevant to baseline scenario
development, including the following: the Low Emissions
Development Strategies Global Partnership (LEDS GP),

the Green Growth Best Practices (GGBP) Initiative, and
the World Bank’s Partnership for Market Readiness
(PMR).

Structure of the report
The report is organised in two parts. Part 1 comprises
this introduction, four analytical chapters and a final
section including reflections by the authors of Part 1.
The analytical chapters cover model choices and uses
(chapter 2), assumptions used in the modelling process
and sensitivity analyses (chapter 3), data management
(chapter 4) and transparency and inclusiveness (chapter
5). Chapter 6 gives the authors’ views on three key issues
related to developing baseline scenarios: good practice,
transparency and uncertainty. Part 2 comprises individual
country experiences as provided by the experts from
each participating country.

5. Available at: />6. Available at: />7. Available at: />8. Available at: o
9. See />10. See />

18

Chapter 2: Model choice and use
In practice, national baseline and mitigation scenarios
are almost exclusively quantitative: they generally rely on
model-derived projections of sectoral activity and sinks,
underpinned by assumptions about GDP, population and
energy prices, among others. The models used and the
assumptions made to prepare those projections have

a strong influence on the resulting scenarios. The main
sectors for GHG emissions in most baseline scenarios
are: energy, agriculture, land-use, industrial processes
and waste. The energy sector and the land-use sector
account for the bulk of GHG emissions in many developing countries. Emissions in the energy sector come
mostly from electricity generation, space heating, industry
and transportation. Land-use sector emissions and sinks
include those resulting from changes to the use of land
(for example, agricultural land converted to urban use);
planting, cutting down or management of forests; and
emissions from the soil.

Types and use of models
Models used to generate projections of GHG emissions
are typically categorised as top-down or bottom-up; the
former approach focuses on economic inter-linkages,
while the latter involves more detailed treatment of
specific technologies (Table 1). Hybrid models, such as
the International Energy Agency’s World Energy Model
(WEM), attempt to bridge the differences between topdown and bottom-up approaches.
In its simplest form, a top-down scenario of energyrelated GHG emissions relies on projections of both future
economic output and overall emissions intensity (defined
as GHG emissions per unit of GDP). The product of these
two series of values over a future time period provides
an anticipated baseline for energy-related emissions (the
model used to generate such a scenario is referred to as
a simple extrapolation model in Table 1).11 More complex
top-down models, such as computable general equilibrium (CGE) models, can simulate interactions among
economic sectors, taking into account their overall effects
on key macroeconomic variables such as consumption,

investment and GDP.

11. This is a simplified version of the Kaya identity which states that the total GHG emissions is the product of four inputs: population, GDP per
capita, energy consumption per GDP and GHG emissions per unit of energy consumed.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

19

Table 1: Overview of model types
Bottom-up

Top-down

Hybrid

Accounting

Optimisation

Simple
extrapolation

Computable
general
equilibrium

Strengths

Ease-of-use and
potentially small
data needs

Technological
detail and leastcost projections

Ease-of-use and
potentially small
data needs

Feed-back
effects on
macroeconomic
variables

Weaknesses

Linkages with broader macroeconomic developments missing

Lack of technological detail

Can be very
resource-intensive

Examples12

LEAP13, MEDEE
and MAED

Spreadsheet
models

WEM (IEA), NEMS,
MARKAL-MACRO
and IPAC

MARKAL/
TIMES, POLES,
RESGEN and
EFOM

Bottom-up models use highly disaggregated data on
specific technologies, such as for energy supply, including estimated costs. This approach makes it possible
to produce fairly detailed projections of energy use by
type and sector, based on assumptions about underlying
drivers such as demographic changes and variations in
consumer income. However, including this level of detail
usually means there is a less thorough characterisation of
the interactions among economic sectors, which are only
represented indirectly through exogenous energy prices,
discount rates and technology learning rates. Bottom-up
models can be sub-divided into accounting models (such
as LEAP) and optimisation models (such as MARKAL/
TIMES). The former allows users to systematically analyse
an assumed structural or policy-related development in
each sector, whereas the latter incorporates some form
of optimising behaviour for economic agents. Up to now,
most national GHG emissions scenarios have relied on
some form of bottom-up model, especially in the case of

energy-related emissions.

ENV-Linkages
(OECD), SGM
and CETA

”

Technological
detail and consistency with economic
projections

By using a CGE-type model in IPAC,
national level fiscal policies including
carbon tax, energy pricing, subsidies
and emissions caps can be analysed.
Similarly, IPAC’s bottom-up technology model can analyse energy
efficiency polices… This capability is
quite similar to that of other modelling
teams in China.
China (ERI)

Hybrid models attempt to combine the advantages of
top-down and bottom-up modelling by linking the two
types of approaches. The main challenge lies in the
complexity of making two models (fundamentally different
in their constructions) run in a consistent manner, which
can require a lot of resources (especially in terms of data
needs) and expertise.

12. Some of these models are proprietary and may not be available for wider use (e.g. WEM); others have been designed specifically to be adapted
and used by third parties (e.g. LEAP).
13. A recent addition to the LEAP model allows for simplified optimisation.


20

Country Experiences
Practices in the ten participating countries span the full
spectrum of modelling approaches, ranging from simple
extrapolation to advanced engineering models (Table 2).

broader macro-economic developments to be taken into
account. Several other hybrid models have also been
used in China.

Most countries rely on bottom-up models (LEAP,
MARKAL/TIMES, MESSAGE/MEAD or purpose-developed models). The appeal of those models lies in their
ability to provide a reasonably detailed representation of
the energy system (which in most countries is the principal source of emissions), while keeping resource needs
down to a reasonable level.

Ethiopia relies on a combination of simplified top-down
and simplified bottom-up modelling. The top-down model
generates projections of broad emission trends, while the
bottom-up model is used to produce additional detail at
the sectoral level.

In China, ERI’s IPAC model is a type of hybrid model,
essentially combining three different models: an emissions model, a technology model and a CGE model. This

design allows the interactions of the energy sector with

The requirements of hybrid models, in terms of both data
and expertise, seem to make them unsuitable for most
participating countries at present. Conversely, simple
top-down models provide a solution for countries with
few resources. Bottom-up models are clearly the tool of
choice for most countries participating in this study.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

21

In practice, the choice of model tends to reflect a tradeoff between model performance and the expected use of
model outputs on the one hand, and resource and data
availability on the other. Performance is often a function of
both the level of sophistication of the model and its suitability to national conditions. Resource constraints take
the form of limits on funding and the technical capacity
within the government departments tasked with preparing baseline and mitigation scenarios.
Resource constraints have been highlighted as a key
factor influencing the choice of model in many of the
participating countries. In Indonesia, this is made more
challenging by a relatively decentralised government
structure, where sub-optimally equipped provincial
entities play a significant role in baseline development.
In such settings, LEAP – a widely-used software tool
for energy policy analysis and climate change mitigation
assessment developed at the Stockholm Environment
Institute – is often the preferred solution. China, Brazil and

South Africa have used more sophisticated bottom-up
and hybrid models, reflecting their longer experience of
modelling and their greater in-country capacity compared
to many other developing countries.
Few estimates exist of the full financial costs incurred
in the preparation of a given baseline scenario, mainly
because of the difficulty in coming up with a reliable
estimate. One reason for this is that modelling tools and
skills are developed and applied gradually, making it hard
to allocate costs to the preparation of a single baseline
scenario.
Nonetheless, the costs can clearly be high relative to
national income in some developing countries. For this
reason, several developed countries have provided technical and financial support for the preparation of baseline
scenarios in developing countries. In addition to easing
the financial burden of preparing the scenarios, this support has also influenced the choice of model, by allowing
countries to opt for more sophisticated models and, in
some instances, because donors may have indirectly
favoured a particular modelling approach (as mentioned
specifically by Vietnam).

”
”
”

The business-as-usual emissions
level for all sectors was developed
using the bottom-up LEAP because
of its flexible data structure, past
experience, transparency and accessibility.

Thailand
The costs of developing the baseline
[is a challenge because it is] fairly
expensive to conduct coordination process and intensive capacity
building for all the local government
officers.
Indonesia
It took two Senior Researchers,
together with several other ERC staff
members, all new to MARKAL, a period of more than a year to complete
the model…
South Africa (ERC)


22

Existing versus purpose-made models
Most developing countries use an existing model to build
their energy-sector emissions scenarios, but some –
most commonly those with especially large or complex
economic and energy systems – develop models customised to their own particular national circumstances.
Some other countries adapt an existing model to their
specific context or combine it with some additional
customised modelling. The choice of which model to use
depends on each country’s institutional capacity, as well
as its particular needs for, and expectations from, the
resulting emissions scenarios.

Several countries have indicated that the choice of
model is influenced by each model’s ease of use and

by the familiarity that governments have with any given
type of model. Once a first baseline scenario has been
prepared with a particular model, there is often interest
in also using that model for subsequent updates, rather
than developing the capacity from scratch to adopt new
modelling tools. This familiarity also helps to give others
in government and in the private sector confidence in the
modelling results.

Country Experiences
Indonesia, Thailand and Vietnam all rely on LEAP for
developing their emissions scenarios. Reasons for this
include ease of use and manageable data requirements.
India (TERI) and South Africa (ERC) both use MARKAL/
TIMES. A convenient user interface and the model’s
optimisation routines are unanimously cited as the main
reasons for this choice.
In Brazil, MESSAGE/MEAD was chosen largely because
key stakeholders, not least the technical agencies
charged to support the baseline development process,
were already familiar with it. This helped to reduce startup costs and ensured broad support for the results.
In Mexico, both the original baseline scenario in 2009
(using a top-down approach) and the revised baseline in
The models used by several of the participating countries
are characterised by a degree of customisation, but only
one country (Mexico) used a fully purpose-made model.
However, this is about to change, as a new update of
the Mexican baseline scenario is currently being finalised
using LEAP. It would appear, therefore, that in most
countries, for fairly homogeneous sectors such as power

2010 (using a bottom-up approach) were prepared using
purpose-made models.
Ethiopia’s approach – a combination of top-down and
bottom-up modelling – was driven by the time and capacity constraints under which the baseline development
process took place. A more sophisticated approach is
envisaged for the future. Kenya also suffered from capacity constraints and opted for a similar simplified approach.
China has used several different models over the years
(see Country Experiences above) to take account of the
interactions of the energy sector with broader macroeconomic developments. ERI’s IPAC modelling team and
several universities in the country use this approach.

generation and also energy-intensive industries such
as cement or iron and steel, generic models provide a
more convenient solution than purpose-made models.
Conversely, modelling of emissions from more diverse
and/or uncommon sectors often relies on custom-made
models, because few, if any, generic off-the-shelf models
are available for those sectors.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

”

23

The COMAP model
is appropriate to the
national circumstances

of Vietnam and depends
on the interest of donors.
Vietnam also has experience with this model,
from the Initial National
Communication of
Vietnam to the UNFCCC.
Vietnam

Land-use sector emissions modelling
The importance of land-use sector emissions varies
significantly from one country to another. While it is a key
source of emissions in Brazil and Indonesia, for example,
the sector makes a very small contribution to overall
emissions levels in the other participating countries.
Modelling approaches range from relatively complex
sector-specific models to simple add-ons to energysector models. These models typically include agriculture, though a separate model is used for agriculture in
Indonesia.

While land-use sector emissions may also be projected
using a top-down model, bottom-up approaches are the
norm in countries where emissions from these sectors
are small or where their economic output is modest. This
is because the expected change in national output over
time may not be a good indicator of the rate of change of
land-use sector emissions, especially in countries where
agriculture and forestry represent only a small share of
economic activity. Established models for projecting
land-use sector emissions and sinks do exist – including
some add-ons to energy sector models – but are less
well-established than energy and emissions models.

Country experiences
Brazil relies on extrapolations of past deforestation trends.
More detailed information from existing satellite observation programmes are being used for planning purposes,
but not for preparing the country’s baseline scenario.
Mexico has integrated land-use-change data into a larger
purpose-made bottom-up model. Conversely, Ethiopia
and Kenya use simple top-down extrapolation methods,
which rely on land-use-change data. Given the varying
quality of these data and the complexity of land-based
emissions modelling, the robustness of those extrapolation methods is similarly variable.

use planning for Low Emission Development Strategy
(LUWES) decision-support framework to develop a
national forestry plan. The plan includes future land uses,
which forms the main set of assumptions for the baseline
scenario. Building on existing work, South Africa has
developed a spreadsheet-based optimisation model for
afforestation (costs included forest establishment, tending, protection, harvesting, transport, overheads and the
opportunity cost of land and water). Vietnam has been
using a pre-existing model (the Comprehensive Mitigation
Analysis Process, or COMAP, model), which had been
used for the preparation of the country’s first national
communication to the UNFCCC.

Indonesia, South Africa and Vietnam rely on more
sophisticated approaches. Indonesia has used the Land

One might expect that countries whose land-use sector emissions account for a large proportion of national
emissions would have a stronger interest in investing in

building modelling capacity in this area. However, experience suggests that existing tools and processes, as
well as resource constraints, are the main determinants

of the sophistication of the modelling approach used.
One reason for this may be the inherent uncertainty that
charcaterises the modelling of emissons from forestry
and land-use-change: beyond a certain level of complexity, the incremental effort needed to enhance the output
appears to be significant.


24

Institutional arrangements and capacity
constraints
Institutional arrangements and the technical expertise and
resources available also influence the choice of method
and approach to preparing a baseline scenario. The way
in which government agencies and, in some cases, academic or other non-governmental entities share responsibility for the task, including the types of co-operation
mechanism to facilitate the exchange of information,
data, and decision-making, differs greatly from country to
country. International co-operation also varies. The existence of a specific political mandate or other formal goals
for baseline scenarios, which may call for the construction
of several baselines based on different assumptions, can
also influence the choice of method.

Irrespective of the chosen modelling tools, the institutional
needs for producing baseline and mitigation scenarios are
large: it generally takes several years for a government
agency to develop all the required tools and build all the
necessary capacities to be able to produce such scenarios with a certain level of sophistication. As capacities

expand, the range of modelling tools may also grow; this
may improve the robustness of the resulting scenarios,
but adds complexity to the process (in particular as
regards the land-use sector) and puts added strain on
already limited budgets and capacities.

Country experiences
The preparation of baseline scenarios is always embedded in broader climate change planning efforts. A variety
of institutional arrangements are used to oversee these
efforts, ranging from formal inter-ministerial committees to
more ad-hoc structures.
In Ethiopia the process of developing the baseline is
part of the Climate Resilient Green Economy Strategy, a
high-profile initiative implemented by the national environmental and development authorities. In South Africa, the
baseline has been developed in support of the country’s
Long Term Mitigation Scenarios process, carried out by a
research team overseen by the Ministry of Environment.

Baseline scenarios are not an end in themselves: they
support broader national and international processes.
As a result, the process of setting baseline scenarios
is inevitably governed by the institutional arrangements
put in place to implement those broader processes.
These arrangements may have been designed with other
purposes in mind and so may not be best adapted to the

In Brazil and Thailand, the development of the baseline
scenario supports national reporting to the UNFCCC,
whereas in Mexico it informed the national climate
change plan. In all three countries, an inter-ministerial

committee was tasked to guide the work. This approach
helped secure support from the ministries concerned and
facilitated the exchange of data between government
departments.
In Vietnam, the environmental authorities prepare the national baseline scenario, coordinating inputs from several
agencies. No formal institutional structure exists, which
has hampered coordination.

task of preparing a baseline scenario. Increased awareness about the importance of baselines, coupled with
stronger political mandates, and increased experience
and resources, could help improve governance arrangements and enhance inter-agency cooperation within
governments in this regard.


National Greenhouse Gas Emissions Baseline Scenarios: Learning from Experiences in Developing Countries

25

Table 2: Overview of the sectors included in baseline scenarios and the models used
Energy

LULUCF

Brazil (UFRJ)

Bottom-up
(MESSAGE/
MAED)

Simple extrapolation of

historical annual
deforestation

China (ERI)

Hybrid model
(IPAC)

Ethiopia

Top-down (simple
extrapolation using spreadsheets)
and bottom-up
(MAC curves)

India (TERI)

Bottom-up
(MARKAL/TIMES)
and CGE models

Indonesia

Bottom-up
(LEAP) for both
provincial and
national level

Kenya

Bottom-up (intensity extrapolation)

Mexico

Bottom-up
(in-house).
Planned future
work: bottom-up
(LEAP)

South Africa
(ERC)

Bottom-up
(MARKAL/TIMES)
and CGE-model

Thailand

Bottom-up
(LEAP)

Vietnam

Bottom-up
(LEAP)

Agriculture

Industrial

Processes

Waste

Included in
energy modelling
LUWES/Abacus
– spatial planning
approach

Included in
LULUCF
modelling

Included in
energy modelling

Simple linear
projection model

Spreadsheet
model

Spreadsheet
model

Spreadsheet
model

Spreadsheet

model

COMAP

Based on
IPCC guidelines

Note: The colours indicate whether sectors are included or not in the baseline scenario (where information was made
available). Green=included, dark grey=not included and light grey=information not provided.
Source: Country contributions (see Part 2).